Taxonomy and Biogeography of Bunopus Spatalurus (Reptilia; Gekkonidae) from the Arabian Peninsula

Home , Agamura, Bunopus, Bunopus tuberculatus, Crossobamon, Crossobamon orientalis, Omanosaura jayakari

1Institute of Evolutionary Biology (CSIC-Universitat Pompeu Fabra), Barcelona, Spain; 2Faculty of Life Sciences and Engineering, Departament Produccio Animal (Fauna Silvestre), Universitat de Lleida, Lleida, Spain; 3CIBIO, Centro de Investigacßao~ em Biodiversidade e Recursos Geneticos, Universidade do Porto, InBio Laboratorio Associado, Vairao,~ Portugal; 4Departamento de Biologia, Faculdade de Ciencias da Universidade do Porto, Porto, Portugal; 5Department of Biology, Villanova University, Villanova PA, USA; 6Department of Ecology, Faculty of Science, Charles University in Prague, Prague, Czech Republic; 7Drıtec 65, Drıtec, Czech Republic; 8Department of Zoology, National Museum, Prague, Czech Republic Taxonomy and biogeography of Bunopus spatalurus (Reptilia; Gekkonidae) from the Arabian Peninsula

1,2 1,3,4 5,† 6 6 PHILIP DE POUS ,LUIS MACHADO ,MARGARITA METALLINOU ,JAN CERVENKA ,LUKA S KRATOCHVIL ,NEFELI 1 7 8 1 2 1 PASCHOU ,TOMA S MAZUCH ,JIRI SMID ,MARC SIMO -RIUDALBAS ,DELFI SANUY and SALVADOR CARRANZA

Abstract In the last decade, taxonomic studies have drastically increased the number of species known to inhabit the Arabian deserts. While ongoing phylogenetic studies continue to identify new species and high levels of intraspecific genetic diversity, few studies have yet explored the biogeographic patterns in this arid region using an integrative approach. In the present work, we apply different phylogenetic methods to infer relationships within the Palearctic naked-toed geckos. We specifically address for the first time the taxonomy and biogeography of Bunopus spatalurus Anderson, 1901, from Arabia using multilocus concatenated and species tree phylogenies, haplotype networks and morphology. We also use species distribution modelling and phylogeographic interpolation to explore the phylogeographic structure of Bunopus spatalurus hajarensis in the Hajar Mountains and the roles of climatic stability and possible biogeographic barriers on lineage occurrence and contact zones in this arid mountain endemism hot spot. According to the inferred topology recovered using concatenated and species tree methods, the genus ‘Bunopus’ is polyphyletic. Bunopus tuberculatus and B. blanfordii form a highly supported clade closely related to Crossobamon orientalis, while the two subspecies of ‘Bunopus’ spatalurus branch together as an independent highly supported clade that diverged during the Miocene according to our estimations. Within B. s. hajarensis, three geographically structured clades can be recognized that according to our estimations diverged during the Late Miocene to Pliocene. The paleodistribution models indicate climatic stability during the Late Pleistocene and the lineage occurrence, and predicted contact zones obtained from phylogeographic interpolation therefore probably result from the older splits of the groups when these lineages originated in allopatry. As demonstrated by the results of the multilocus molecular phylogenetic analyses and the topological test carried out in this study, the genus ‘Bunopus’ is not monophyletic. To resolve this, we resurrect the genus Trachydactylus Haas and Battersby, 1959; for the species formerly referred to as Bunopus spatalurus. Considering the morphological differences, the high level of genetic differentiation in the 12S mitochondrial gene and the results of the phylogenetic and the cmos haplotype network analysis, we elevate Trachydactylus spatalurus hajarensis to the species level Trachydactylus hajarensis (Arnold, 1980).

Key words: Paleodistribution modelling – phylogeography – multilocus phylogeny – spatial interpolation – contact zone – Palearctic naked-toed gecko

Introduction identify new species, especially in the southern Arabian Penin- sula (Babocsay 2004; Busais and Joger 2011; Carranza and Deserts encompass a large portion (12%) of the Earth’s land sur- Arnold 2012; Metallinou and Carranza 2013; Sm ıd et al. 2013b, face and are important for understanding global biodiversity pat- 2015; Vasconcelos and Carranza 2014) as predicted by Ficetola terns. While deserts are thought to have relatively low species et al. (2013). Likewise, several studies have revealed high levels richness compared to other biomes, they are often inhabited by of intraspeciﬁc genetic structure of Arabian lizards, mostly in many specialized species and clades possessing a wide array of species inhabiting mountainous areas such as the Hajar Moun- unique adaptations to arid conditions. For example, lizard com- tains in northern Oman and the UAE and the mountains in munities in deserts might be richer compared to warm temperate southern Oman and Yemen (e.g. Carranza and Arnold 2012; and tropical regions, and they can contain as many as 70 differ- Sm ıd et al. 2013a), but also in species occurring in lowlands ent species co-occurring at single localities (Pianka 1973; (Metallinou et al. 2012, 2015). Two biodiversity-rich areas with Rabosky et al. 2011). In the past, extensive work on desert rep- high levels of endemicity are recognized within Oman: the Hajar tiles has mainly focused on Australian, South African and North Mountains in north Oman and the Dhofar Mountains in south American deserts (e.g. Pianka 1986), while fewer studies exist Oman and east Yemen. These two mountainous regions have for the Arabian deserts (e.g. Anderson 1896; Haas 1957; Arnold their own unique and complex geologic histories (reviewed in 1972, 1980, 1986; Gasperetti 1988; Sch€atti and Gasperetti 1994; Carranza and Arnold 2012 and Gardner 2013; see also Arnold Sch€atti and Desvoignes 1999; Carranza and Arnold 2012; Gard- 1977, 1980 and other articles in the same volumes) but share dis- ner 2013). Recently, however, an increasing number of taxo- tinct climatic conditions and vegetation that differentiate them nomic, biogeographic and phylogenetic studies have focused on from the much more arid surrounding lowland desert. Although lizards inhabiting these deserts (e.g. Carranza and Arnold 2012; the Hajar and the Dhofar Mountains are inhabited by partially Metallinou et al. 2012, 2015; Metallinou and Carranza 2013; distinct reptile communities with unique endemic species (e.g. Sm ıd et al. 2013a). In the last decade, taxonomic studies have Hemidactylus, Asaccus geckos) and even genera (Omanosaura, drastically increased the number of species inhabiting the Lacertidae) (see Arnold 1986; Arnold and Gardner 1994; Car- Arabian desert, while ongoing phylogenetic studies continue to ranza and Arnold 2012), several species occur in both regions while being absent from the lowland areas between them (Buno- Corresponding author: Salvador Carranza ([email protected]) pus spatalurus Anderson, 1901, Chalcides ocellatus (Forskal, †Author deceased July 2015

J Zool Syst Evol Res (2016) 54(1), 67--81 68 DE POUS,MACHADO,METALLINOU, CERVENKA,KRATOCHVIL,PASCHOU,MAZUCH, SMID,SIMO-RIUDALBAS,SANUY and CARRANZA

1775), Trachylepis tessellata (Anderson, 1895) and Platyceps the genus Bunopus (Bauer et al. 2013), as well as three of the four spe- rhodorachis (Jan, 1865); see Gardner 2013). cies of this genus (Bunopus crassicauda Nikolsky, 1907 missing) and Among the species exhibiting this notable north–south distri- two out-groups (Hemidactylus brasilianus (Amaral, 1935) and bution pattern, Bunopus spatalurus is an interesting representa- H. haitianus Meerwarth, 1901). Given the focus of the study, sampling tive, as previously mentioned by Arnold (1980). This species is was especially intensive in Bunopus spatalurus. A list of all the specimens with their GenBank accession numbers, voucher codes and locality widely distributed in the southern Arabian Peninsula, ranging information is presented in Table 1. from western Yemen to the mountains of northern Oman and the Genomic DNA was extracted from ethanol-preserved tissue samples UAE, but with some areas of absence (Sindaco and Jeremcenko using the standard high salt method (Sambrook et al. 1989). Up to six 2008; Gardner 2013). Arnold (1980) described the subspecies genetic markers were PCR-amplified and sequenced in both directions Bunopus spatalurus hajarensis from the Hajar Mountain range (see Table S1 for primer details): one fragment of the mitochondrial gene and Masirah Island based on clear morphological differences. encoding the ribosomal 12S rRNA (12S; primers 12SaGekko and However, despite some recent phylogenetic studies on Palearctic 12SbGekko – Metallinou et al. 2015), and five fragments of the nuclear naked-toed geckos (Cervenka et al. 2008; Bauer et al. 2013), no genes encoding the oocyte maturation factor Mos (cmos; primers FU-F – molecular studies have yet addressed the status of Bunopus and FU-R Gamble et al. 2008), the recombination-activating gene 1 – – spatalurus. This is possibly the result of the difficulty in finding (rag1; primers F700, R700 Bauer et al. 2007; and R13 and R18 Groth and Barrowclough 1999), the recombination-activating gene 2 specimens of B. s. spatalurus compared with the relatively abun- (rag2; primers PY1-F and PY1-R – Gamble et al. 2008), the acetyl- dant B. s. hajarensis (Haas and Battersby 1959; Arnold 1980; cholinergic receptor M4 (acm4; primers int-F and int-R – Gamble et al. Gardner 2013) as well as the current political instability in 2008) and a short fragment of phosducine (pdc; primers PHOF2 and Yemen that obstructs fieldwork. Clarifying the phylogenetic posi- PHOR1 – Bauer et al. 2007). PCR conditions used for the amplification tion of B. s. spatalurus and B. s. hajarensis should provide of the 12S mitochondrial fragment are as in Metallinou et al. (2015), for important insights into Arabian reptile biogeography in general, nuclear genes cmos, rag1 and rag2 as in Sm ıd et al. (2013a), for the while it would also contribute to clarifying the taxonomy of the nuclear gene acm4 in Barata et al. (2012) and for pdc in Bauer et al. genus Bunopus Blanford, 1874. (2007). The geologically complex Hajar Mountains in northern Oman and the UAE are known for their numerous endemics, and recent Sequence analysis studies have identified high genetic variation of reptile species inhabiting this massif (Papenfuss et al. 2010; Carranza and GENEIOUS v. R6.1.6 (Biomatters Ltd., Auckland, New Zealand) was used Arnold 2012). However, no studies have explored the possible for assembling and editing the chromatographs. Heterozygous positions fi causes of this genetic structuring or investigated the historical for the nuclear coding gene fragments were identi ed based on the presence of two peaks of approximately equal height at a single nucleotide biogeography of this arid region in detail using an integrative site in both strands and were coded using IUPAC ambiguity codes. The approach. Paleodistribution modelling using climatic data sets nuclear coding fragments were translated into amino acids and no stop from the Last Glacial Maximum and the Mid-Holocene can help codons were observed. DNA sequences were aligned for each gene inde- elucidate the roles of climate oscillations and stability and has pendently using the online application of MAFFT v.7 (Katoh and Standley been used successfully in other interesting biodiversity hot spots 2013) with default parameters (auto strategy, gap opening penalty: 1.53, (e.g. Waltari et al. 2007; Carnaval et al. 2009). Furthermore, offset value: 0.0). For the 12S ribosomal fragment, the Q-INS-i strategy promising methodological advances in spatial phylogeographic was applied, in which information on the secondary structure of the RNA interpolation (e.g. Tarroso et al. 2015) further expand the possi- is considered. Poorly aligned regions in the 12S alignment were elimi- bilities to explore lineage occurrence and identify contact zones nated with Gblocks (Castresana 2000) under low stringency options cmos between lineages. Such spatial modelling methods should pro- (Talavera and Castresana 2007). In the case of the gene used in the network analyses, the software PHASE v. 2.1.1 was used to resolve vide interesting new insights into the roles of climatic fluctua- heterozygous positions (Stephens et al. 2001), and SEQPHASE (Flot 2010) tions and refugia on phylogeographic structure in this high was used to convert the input and output files. Default settings of PHASE endemism area. were used except for phase probabilities that were set as ≥0.7 (see Harri- In the present work, we revise the systematics of the genus gan et al. 2008). Uncorrected mean genetic distances between and within Bunopus using multilocus concatenated and species tree phyloge- groups for the mitochondrial gene fragment were calculated with MEGA 5 nies, haplotype networks and morphology. We also use species (Tamura et al. 2011), using the p-distance model. distribution modelling and phylogeographic interpolation to explore the phylogeographic structure of Bunopus spatalurus in Phylogenetic and network analyses and estimation of the Hajar Mountains and the roles of climatic stability and possi- divergence times ble biogeographic barriers on lineage occurrence and contact zones in this arid mountain endemism hot spot. Two data sets were used for the phylogenetic analyses. Data set 1 was We considered the status of the Operational Taxonomic Unit assembled with the aim of testing the monophyly of the genus Bunopus and its relationships with the most closely related genera of naked-toed (OTUs) in this study within the framework of the general lineage species concept (de Queiroz 1998, 2007), which uses the term geckos (Cervenka et al. 2008; Bauer et al. 2013). This data set consisted ‘species’ for separately evolving metapopulation lineages. We of 12 representatives of the Palearctic naked-toed geckos of the genera Agamura Blanford, 1874, Crossobamon Boettger, 1888, Cyrtopodion Fit- apply two secondary recognition criteria to assess species limits zinger, 1843, Tenuidactylus Szczerbak and Golubev, 1984 and Bunopus within Bunopus spatalurus using the following approach: (1) and two representatives of the genus Hemidactylus Oken, 1817 selected identification of lineages based on the analyses of mitochondrial as out-groups based on Gamble et al. (2012) and Bauer et al. (2013). We and nuclear data as indicators of distinct evolutionary trajectories included members of all the species and subspecies of Bunopus with the and (2) the presence of diagnostic morphological characters. only exception of B. crassicauda Nikolsky, 1907, which is closely related to Bunopus tuberculatus Blanford, 1874 sensu lato (Cervenka et al. 2008). Data set 1 consisted of an alignment of 3020 base pairs (bp) Materials and Methods of concatenated mitochondrial (379 bp of 12S) and nuclear (742 bp of Molecular samples, DNA extraction and amplification cmos; 1038 of rag1; 363 of rag2; 444 bp of acm4; 394 of pdc) DNA sequences. Data set 2 was assembled with the aim of studying in detail A total of 42 specimens were used in the phylogenetic analyses, includ- the phylogeographic relationships of the two endemic Arabian subspecies ing members of the Palearctic naked-toed geckos most closely related to of the genus Bunopus (Bunopus s. spatalurus and Bunopus s. hajarensis)

Table 1. Specimens of the naked-toed geckos included in the molecular analyses. For all 34 specimens newly sequenced in the present study we provide their taxonomic identiﬁcation, sample code, voucher reference, corresponding geographical distribution data (latitude, longitude and country), locality code as shown in the map from Fig. 1 and GenBank accession numbers. For the specimens not sequenced in the present study we only provide the GenBank accession numbers for the genes used and in one case that all the genes correspond to the same specimen we also provide the country.

Species Sample code Voucher Lat. Long. Country Locality code 12S cmos rag1 rag2 acm4 pdc

Agamura persica DQ852726 JQ945528 JQ945281 JQ945420 JQ945634 JQ945349 Bunopos blanfordii J27 31.583 37.25 Jordan KT302094 KT302127 KT302139 KT302144 KT302100 KT302129 Bunopus tuberculatus T31 29.633 50.433 Iran KT302095 KT302128 KT302140 KT302145 KT302101 KT302130 Bunopus tuberculatus UAE – JQ945535 JQ945287 JQ945427 JQ945641 JQ945355 Crossobamon orientalis

DQ852715 JQ945547 JQ945299 JQ945440 JQ945653 JQ945368 spatalurus Bunopus Cyrtopodion scabrum EU589172 HQ426532 HQ426275 HQ426448 HQ426355 HQ426186 Hemidactylus brasilianus DQ120428 HQ426523 EU268290 HQ426439 HQ426346 EU268320 Hemidactylus haitianus DQ120388 HQ426543 HM559700 – HQ426368 HM559667 Tenuidactylus caspius EU589164 JQ945620 JQ945340 JQ945514 JQ945727 JQ945409 Tenuidactylus longipes EU589170 JQ945621 JQ945341 JQ945515 JQ945728 JQ945410 Trachydactylus spatalurus SPAT01 NMP 75139/1 13.877 45.800 Yemen 30 KT302090 KT302122 KT302136 – KT302097 KT302132 Trachydactylus spatalurus SPAT02 NMP 75139/2 13.877 45.800 Yemen 30 KT302091 KT302123 – ––– Trachydactylus spatalurus TMHC404 NMP 75147 17.0289 54.6665 Oman 28 KT302092 KT302124 KT302137 KT302143 KT302098 KT302131 Trachydactylus spatalurus TMHC405 NMP 75146 16.8844 53.7731 Oman 29 KT302093 KT302125 – ––– Trachydactylus hajarensis CN4226 22.107 59.357 Oman 24 KT302089 KT302108 – ––– Trachydactylus hajarensis CN3853 22.5401 59.6408 Oman 23 KT302088 KT302126 KT302138 KT302142 KT302099 KT302134 Trachydactylus hajarensis S1755 20.3118 58.7366 Oman 26 KT302087 KT302121 – ––– Trachydactylus hajarensis CN677 20.4981 58.9306 Oman 25 KT302085 KT302109 – ––– Trachydactylus hajarensis S1165 20.2995 58.7497 Oman 27 KT302086 KT302110 – ––– Trachydactylus hajarensis S7643 23.101 57.3496 Oman 15 KT302063 –– ––– Trachydactylus hajarensis AO27 22.905 57.53 Oman 16 KT302062 KT302102 – ––– Trachydactylus hajarensis CN686 IBE CN686 22.9492 59.1983 Oman 17 KT302064 KT302118 – ––– Trachydactylus hajarensis S7161 22.6161 59.0937 Oman 22 KT302067 KT302120 – ––– Trachydactylus hajarensis S7150 23.1317 58.6189 Oman 20 KT302066 KT302119 – ––– Trachydactylus hajarensis CN3750 IBE CN3750 23.0865 57.6762 Oman 17 KT302065 –– ––– Trachydactylus hajarensis OM100 NMP 74270/2 23.2543 57.9315 Oman 19 KT302070 –– ––– Trachydactylus hajarensis OM66 NMP 74268 23.0535 57.805 Oman 18 KT302068 –– ––– Trachydactylus hajarensis OM99 NMP 74270/1 23.2543 57.9315 Oman 19 KT302069 –– ––– Trachydactylus hajarensis CN3989 IBE CN3989 23.377 57.6644 Oman 14 KT302074 KT302113 – ––– olSs vlRs(2016) Res Evol Syst Zool J Trachydactylus hajarensis CN664 23.1498 56.8942 Oman 12 KT302071 KT302103 – ––– Trachydactylus hajarensis CN2575 23.7102 56.4431 Oman 11 KT302073 KT302105 KT302135 KT302141 KT302096 KT302133 Trachydactylus hajarensis CN681 23.22 56.9736 Oman 13 KT302072 KT302104 – –––

© Trachydactylus hajarensis CN2641 24.9936 56.2167 UAE 8 KT302075 KT302107 – ––– – ––– 05BakelVra GmbH Verlag Blackwell 2015 Trachydactylus hajarensis CN3433 24.6208 56.3398 Oman 9 KT302076 KT302111 Trachydactylus hajarensis CN3970 IBE CN3970 26.0421 56.3697 Oman 1 KT302084 KT302112 – ––– Trachydactylus hajarensis CN8706 25.9758 56.1504 Oman 2 KT302083 KT302116 – ––– Trachydactylus hajarensis CN8281 25.1823 56.229 UAE 7 KT302081 –– ––– Trachydactylus hajarensis CN3986 IBE CN3986 25.459 56.1834 UAE 5 KT302078 –– ––– Trachydactylus hajarensis CN8355 25.7863 56.2169 Oman 4 KT302082 KT302114 – ––– Trachydactylus hajarensis CN7658 25.3001 56.0453 UAE 6 KT302079 KT302106 – ––– Trachydactylus hajarensis CN3412 IBE CN3412 24.5131 56.4634 Oman 10 KT302077 KT302117 – ––– 54 Trachydactylus hajarensis CN7819 25.8798 56.2144 Oman 3 KT302080 KT302115 – ––– 1,67--81 (1),

IBE: Institute of Evolutionary Biology, Barcelona, Spain; NMP: National Museum, Prague, Czech Republic; Taxon names correspond to changes proposed in this paper. 69 70 DE POUS,MACHADO,METALLINOU, CERVENKA,KRATOCHVIL,PASCHOU,MAZUCH, SMID,SIMO-RIUDALBAS,SANUY and CARRANZA and to assess their subspecific status. This data set consisted of 32 sam- uniform (0,10). Partitions and clock models were unlinked, and the xml ples, four Bunopus s. spatalurus and 28 Bunopus s. hajarensis, from 30 file was manually modified to set Ambiguities=‘true’ for the nuclear localities across the entire distribution range of the species in Arabia genes in order to account for variability in the heterozygous positions, (Fig. 1). In the face of uncertainty over the sister group to Bunopus instead of treating them as missing data. Posterior trace plots and effec- spatalurus (see Results below) and in order to optimize the alignment of tive sample sizes (ESS) of the runs were monitored in TRACER v1.5 (Ram- the 12S gene, we did not include any out-groups in data set 2 and use baut and Drummond 2007) to ensure convergence. The results of the Bayesian methods for inferring the root of the phylogenetic tree individual runs were combined in LogCombiner discarding 10% of the (Huelsenbeck et al. 2002). Data set 2 consisted of an alignment of samples, and the maximum clade credibility (MCC) ultrametric tree was 396 bp of the 12S gene. produced with TreeAnnotator (both provided with the BEAST package). Best-fitting models of nucleotide evolution for data set 1 were inferred Phylogenetic analyses for data set 2 were performed with BI only, with using PARTITIONFINDER v.1.1.1 (Lanfear et al. 2012) with the following set- the same parameters as above but using a coalescent constant size pro- tings: branch lengths linked, only models available in BEAST evaluated, cess tree prior. Nodes were considered strongly supported if they received AIC model selection criterion applied, and all partition schemes analysed. posterior probability (pp) support values ≥0.95 (Wilcox et al. 2002; Each gene was set as an independent partition. A two-partition scheme Huelsenbeck and Rannala 2004). was selected: p1, 12S gene and the GTR+G model of sequence evolution; The lack of internal calibration points in the Palearctic naked-toed p2, all five nuclear genes and the HKY+G model of sequence evolution. geckos precluded the direct estimation of the time of the cladogenetic For data set 2, we used JMODELTEST v.0.1.1 (Guindon and Gascuel 2003; events in our phylogenies. Alternatively, the substitution rate of the same Darriba et al. 2012) under the Akaike information criterion (AIC) mitochondrial region calculated for other lizard groups could be used for (Akaike 1973) and the HKY+I+G model of sequence evolution was this purpose. Mean substitution rates and their standard errors for the selected. For both data sets, we performed for each partition a likelihood same 12S gene regions used in this study were extracted from fully cali- ratio test implemented in MEGA v.5 (Tamura et al. 2011) to test whether a brated phylogenies including various lizard groups from the Canary strict or a relaxed molecular clock fit our data best. The hypothesis that islands: Tarentola sp. (Phyllodactylidae) (Carranza et al. 2000, 2002), the sequences evolve in a clock-like manner could not be rejected at a Gallotia sp. (Lacertidae) (Cox et al. 2010), and Chalcides sp. (Scincidae) 5% significance level for all nuclear genes of data set 1, while it was (Brown and Pestano 1998; Carranza et al. 2008; Brown and Yang 2010) rejected for the 12S gene in both data sets. and Podarcis pityusensis (Bosca, 1883) and P. lilfordi (Gunther,€ 1874) Phylogenetic analyses for data set 1 were performed with maximum- (Lacertidae) from the Balearic islands (Brown et al. 2008). For a full likelihood (ML) and Bayesian inference (BI) methods. ML analyses were account on the specific calibration points and methods used to infer the performed in RAXML v.7.4.2 (Stamatakis 2006) as implemented in raxml- individual substitution rates for each lizard group and the combined sub- GUI (Silvestro and Michalak 2012) with 100 random addition replicates, stitution rate used in this study, please see Carranza and Arnold (2012). using the GTR+G model of sequence evolution and independent model Despite the problems associated with using evolutionary rates from other parameters for the two partitions. Reliability of the ML tree was assessed organisms for time tree calibration, the rate of the 12S rRNA gene by bootstrap analysis (Felsenstein 1985) including 1000 replications. The inferred by Carranza and Arnold (2012) and applied here has been cor- software BEAST v.1.7.5 (Drummond and Rambaut 2007) was used for BI roborated by independent studies that used different calibration points analysis of the concatenated data set as well as species tree analysis and different taxa to the ones employed by Carranza and Arnold (2012) employing *BEAST (Heled and Drummond 2010). Three individual runs (Metallinou et al. 2012; Sindaco et al. 2012). Moreover, the rates by Car- of 5 9 107 generations were carried out, sampling at intervals of 10 000 ranza and Arnold (2012) have already been applied to several different generations. Models and prior specifications applied were as follows studies of lizards for which reliable internal calibration points based on (otherwise by default): model of sequence evolution for each of the two geographical or fossils evidence do not exist (Hawlitschek and Glaw partitions of data set 1, as indicated above; Yule process tree prior; ran- 2013; Mila et al. 2013; Smıd et al. 2013a; Vasconcelos and Carranza dom starting tree; base substitution prior uniform (0,100); and alpha prior 2014; Metallinou et al. 2015). As a result of that, absolute divergence

Fig. 1. Map of localities of material examined in this study. Circles indicate samples used in the genetic analyses (red = Trachydactylus spatalurus, blue = T. hajarensis), with different shades of blue representing the distinct clades of T. hajarensis as shown in Fig. 2. A list of all the specimens with their taxonomic identiﬁcation, sample code, voucher code, country, and corresponding geographical distribution data and GenBank accession numbers for the sequenced genes is presented in Table 1. Taxon names correspond to changes proposed in this study. J Zool Syst Evol Res (2016) 54(1), 67--81 © 2015 Blackwell Verlag GmbH Taxonomy and biogeography of Bunopus spatalurus 71 times were estimated concurrently in the Bayesian analyses of both data curve plots, which plot the true-positive rate against the false-positive sets, applying Carranza and Arnold (2012) mean rate of molecular evolu- rate. The average area under the curve (AUC) of the ROC plot of ten tion for the same 12S gene fragment (mean: 0.00755, stdev: 0.00247). models was taken as a measure of the overall ﬁt of each model. Compar- A data set including sequences of the nuclear gene cmos for 24 speci- isons of the environmental variables used for projection to those used for mens included in data set 2 (Table 1) was used to infer the genealogical training the model were made using visual interpretation of multivariate relationships within Bunopus spatalurus using a haplotype network similarity surface (MESS) pictures and the most dissimilar variable inferred with statistical parsimony as implemented in the program TCS (MoD) (Elith et al. 2010). v.1.21 (Clement et al. 2000). Phased sequences were used, and a connection limit of 95% was applied. Phylogeographic interpolation

Topology test The recently developed PHYLIN R package (Tarroso et al. 2015) was used to spatially interpolate the phylogeographic data of Bunopus s. hajaren- To assess the polyphyletic status of the genus Bunopus, a tree with the sis. These interpolations can be used to predict the spatial occurrence of alternative topology (topological constraint) in which the members of the different lineages within a phylogeny using a modified method of kriging. genus Bunopus were forced to form a monophyletic group was recon- We used only Bunopus s. hajarensis haplotypes from data set 2 to create structed using data set 1. The constrained topology was compared to the an ultrametric BI tree in BEAST, as above. The same extent and resolution unconstrained, best ML, tree using the approximately unbiased (AU; Shi- as used for the SDM was used as study region for interpolation. A total modaira 2002) and Shimodaira–Hasegawa (SH; Shimodaira and Hase- variogram was produced with the default values (lag = 0.6°, gawa 1999) tests. Per-site log likelihoods were estimated using lagtol = 0.3°), and a model was fitted with sill and range estimated by raxmlGUI, and p-values were calculated using CONSEL (Shimodaira and nonlinear least squares with nugget forced to 0. Additionally, we investi- Hasegawa 2001). gated the spatial dependence of the three main clades (1–3) using cluster variograms with fitted model by nonlinear least squares or forced range size. All variograms were created and fitted with a spherical model. PHY- Species distribution modelling LIN was used to interpolate clade occurrence following the 0.95 probability and to identify potential contact zones between the three main clades A total of 113 distribution records of Bunopus spatalurus hajarensis were recognized in the phylogenetic analyses (see Results) using a single assembled from literature and fieldwork. Spatially autocorrelated distribu- threshold (hs = 6). tion records were removed using a spatial rarefying protocol as implemented in SDMTOOLBOX (Brown 2014). After spatial rarefying, 65 distribution records were retained for species distribution modelling Results (SDM) (Figure S1A and Table S2). Bioclimatic variables were down- Molecular analyses loaded from the WORLDCLIM database version 1.4 (Hijmans et al. 2005) at 9 a resolution of 2.5 arc minutes (nearly 5 5 km). Past climate data for The results of the concatenated ML analysis of the selected the Last Glacial Maximum (LGM) and Mid-Holocene (MH) were Palearctic naked-toed geckos using data set 1 are presented in obtained from the WorldClim database as well at a similar resolution. The average of three (LGM: CCSM4, MIROC-ESM and MPI-ESM) and Fig. 2. The BI concatenated (BEAST) and species tree nine (MH: BCC-CSM1-1, CCSM4, CNRM-CM5, HadGEM2-CC, Had- (*BEAST) analyses supported the same phylogenetic relation- GEM2-ES, IPSL-CM5A-LR, MIROC-ESM, MPI-ESM-P, MRI-CGCM3) ships except for the position of Agamura persica (Dumeril, global climate models was used for each time period, respectively. The 1856), which is inferred as sister taxon to Bunopus spatalurus in data were projected to the Asia South Albers Equal Area Conic projec- the ML analysis and as sister taxon to the other members of the tion and resampled to a resolution of 5 km. Collinearity of the initial naked-toed geckos in the Bayesian analyses, although this rela- variables was measured by Pearson’s correlation coefficient in ENMTOOLS tionship is supported with a high pp value only in the BEAST 1.3 (Warren et al. 2010). A total of five bioclimatic variables, all of analysis (Figs 2 and S2). Diversification in this clade of geckos ’ fi which had a correlation degree lower than 0.75 (Pearson s coef cient), is estimated to have initiated approximately 22.7 Ma (13.9–32.2, were retained. The final set of variables used for the Bunopus s. hajaren- 95% HPD) or 25.3 Ma (15.5–35.6, 95% HPD) based on sis SDMs consisted of temperature seasonality (BIO4), max temperature * of warmest month (BIO5), annual precipitation (BIO12), precipitation of BEAST analyses. According to the topology inferred using con- * driest month (BIO14) and precipitation of warmest quarter (BIO18). catenated (ML and BEAST) and species tree ( BEAST) meth- ‘ ’ Species distribution models were generated using the presence/back- ods, the genus Bunopus is polyphyletic. Bunopus tuberculatus ground algorithm MAXENT, version 3.3.3k (Phillips et al. 2006). The EN- and Bunopus blanfordii Strauch, 1887, form a highly supported MEVAL R package (Muscarella et al. 2014) was used to tune and evaluate clade closely related to Crossobamon orientalis (Blanford, 1876), the SDMs for Bunopus s. hajarensis based on the ‘checkerboard2’ while the two subspecies of ‘Bunopus’ spatalurus cluster method under a wide variety of settings and regularization parameters in together in an independent highly supported clade and originated a criterion-based model selection framework. Competing models were in the Late Miocene, approximately 7.6 Ma (4.4–11.1, 95% compared using the AICc as suggested by Warren and Seifert (2011). HPD) or 7.9 Ma (3.9–12.4, 95% HPD) based on *BEAST analy- Maxent was subsequently used with the following settings: convergence ses. There is also limited support for the monophyly of the group threshold = 0.00001, maximum number of iterations = 500 and bj = 3.0 while using linear, quadratic and hinge features. We followed the sugges- comprising Cyrtopodion scabrum (Heyden, 1827), the two spe- tion of VanDerWal et al. (2009) and used an exploratory analysis to cies of Tenuidactylus and the clade formed by Crossobamon ori- define the most appropriate calibration region. Final models were cali- entalis, B. tuberculatus and B. blanfordii (Figs 2 and S2). In all brated in a background region that encompassed all known localities and the analyses, the specimen of B. tuberculatus from UAE is more included areas that have been accessible to the species via dispersal over closely related to B. blanfordii than to the other specimen of relevant time periods (Merow et al. 2013). The average of ten pseudo- B. tuberculatus from Iran included in the phylogenetic analyses replicated models with randomly selected test samples was used to pro- (Fig. 2 and Table 1). fi duce SDMs, which were plotted in logistic format. The nal models were The results of the topology test carried out with data set 1, in fi – reclassi ed in ARCGIS 10 (ESRI) into the binary presence absence maps which all seven specimens of the genus ‘Bunopus’ included in using the maximum training sensitivity plus specificity threshold the analyses were forced to form a monophyletic group, indi- (MTSPS), which maximizes the sum of sensitivity (proportion of actual fi positives that are correctly identified) and specificity (proportion of nega- cated that the constrained topology was signi cantly different tives that are correctly identified) and has been shown to produce highly from the best ML topology presented in Fig. 2, rejecting the accurate predictions (Liu et al. 2005; Jimenez-Valverde and Lobo 2007). monophyly of the genus ‘Bunopus’ (AU p < 0.00001; SH All models were tested with receiver operating characteristics (ROC) p < 0.00001).

Fig. 2. Maximum-likelihood (ML) tree of the Palearctic naked-toed geckos related to Bunopus and Trachydactylus, inferred using the concatenated data set (12S, cmos, rag1, rag2, acm4, pdc). The two Hemidactylus used to root the tree are not shown. Bootstrap values >70% in the ML analysis are shown next to the corresponding nodes. Empty circles indicate pp > 0.95 in the concatenated (BEAST) and species tree (*BEAST) analyses; black-ﬁlled circle indicates pp >0.95 in the BEAST analysis only. An asterisk by the node (*) of the ML tree indicates that this topological relationship was not supported by the BEAST or *BEAST analyses (see Supporting information Fig. 2). Ages of all well-supported relationships inferred using concatenated BI analysis (BEAST) or species tree analysis (*BEAST, in bold letters) are shown by the nodes in italics with the 95% HPD between brackets. Taxon names correspond to changes proposed in this study.

The BI tree inferred using data set 2 is presented in Fig. 3, the Pleistocene. The uncorrected genetic distances (p-distance) and like in Fig. 2, it shows that the two taxa form two well-sup- for the 12S mitochondrial gene among the 32 samples included ported reciprocally monophyletic groups. Within ‘Bunopus’ s. in data set 2 are presented in Table S3. The level of genetic spatalurus, there are two well-supported clades, one with repre- divergence in the 12S between the two subspecies of ‘Bunopus’ sentatives from Yemen (loc. 30 in Fig. 1) and the other from spatalurus is 13.0 Æ 2%. The genetic divergence within ‘Buno- Dhofar, Oman (locs. 28–29), with a relatively recent divergence, pus’ s. spatalurus is only 2.0 Æ 0.6%, while it is 4.5 Æ 0.7% approximately 2.7 Ma (0.8–8.0, 95% HPD). Within ‘Bunopus’ s. for ‘Bunopus’ s. hajarensis. The genetic divergence between the hajarensis, three geographically structured clades can be recog- three clades of ‘Bunopus’ s. hajarensis is 6.9 Æ 1.2% between nized: clade 1, sister clade to the other two clades and restricted clades 1 and 2; 6.6 Æ 1.2% between clades 1 and 3; and to the north-eastern tip of Oman, in the isolated massifs of the 6.2 Æ 1.1% between clades 2 and 3. The level of genetic vari- Jebel Khamis, Jebel Qahwan and surrounding areas and in ability within each one of the clades is 0.5 Æ 0.3%, 1.7 Æ 0.4% Masirah Island (locs. 23–27 in Fig. 1); clade 2, distributed across and 1.1 Æ 0.3% for clades 1, 2 and 3, respectively. the Eastern Hajars and the Jebel Akhdar (locs. 15–22 in Fig. 1); The nuclear cmos haplotype network analysis shows that the and clade 3, distributed from the western foothills of the Jebel two subspecies of ‘Bunopus’ spatalurus do not share haplotypes Akhdar, across the Western Hajars and up to the northernmost (Fig. 4). The three haplotypes revealed in ‘Bunopus’ s. spatalu- tip of the Hajar mountain range (Musandam Peninsula, Oman) rus are separated by a maximum of two mutations. The high (locs. 1–14 in Fig. 1). Although monophyly of each of these level of genetic variability within ‘Bunopus’ s. hajarensis,as three clades is well supported, their inter-relationships are not observed in the 12S mitochondrial gene, with clades 1–3 diverg- totally resolved (Fig. 3). According to our estimations, diversiﬁ- ing by a p-distance above 6% in all three comparisons, is also cation in this taxon initiated approximately 5.7 Ma (2.4–13.2, apparent in the cmos nuclear gene, for which 12 haplotypes sepa- 95% HPD), and in each of the three clades, it took place during rated by 1–4 mutational steps were recovered (Fig. 4). Despite

Fig. 3. Bayesian inference maximum clade credibility tree of Trachydactylus spatalurus and T. hajarensis inferred using BEAST on data set 2 (12S mtDNA gene). A list of all the specimens with corresponding sample code, voucher code, country, locality code and GenBank accession numbers is presented in Table 1. Ages of some relevant nodes are shown by the nodes in italics with the 95% HPD between brackets. Taxon names correspond to changes proposed in this study. Empty circles indicate pp > 0.95. the high level of genetic variability at both mitochondrial and areas outside the training range. For example, most areas in the nuclear levels and the geographical coherence of the three mito- Musandam Peninsula and the UAE as well as parts of the Jebel chondrial clades of ‘Bunopus’ s. hajarensis across its distribution Akhdar have areas with non-analogue climatic conditions, mainly range, all three clades share cmos nuclear haplotypes. Clades 1 for BIO4 but also BIO5, BIO12 and BIO18 (see Figure S3). and 2 present two private haplotypes each, and clade 3, which is represented in total by six haplotypes, has five private haplo- Phylogeographic interpolation types. The spherical model fitted to the total variogram obtained from PHYLIN had a good fit as suggested by nonlinear least squares Species distribution modelling (R2 = 0.957) (Figure S4). The model indicates an isolation-by- Maxent produced SDMs of good predictive accuracy according distance pattern, with small genetic differences at short distances to the average test AUC (0.796 Æ 0.063). The present and past and stabilization of the semi-variance at larger distances. The SDMs for ‘Bunopus’ spatalurus hajarensis reveal large suitable cluster variograms for clades 1–3, however, indicate only a good areas in the Hajar Mountains and surrounding areas that have fit for clade 3 (R2 = 0.8540) and a very poor fit of clades 1 and remained stable and extend well beyond the current distribution 2(R2 = NA and R2 = 0.096). The maps of predicted occurrence range (Figure S1). Although the SDMs predict stability in all of each lineage predicted well the spatial pattern of the three areas in the Hajar Mountains, there is a large area of unsuitable clades in north Oman, despite the low fit models of clades 1 and climate between the Hajars and Masirah Island in the Sharqiyah 2 (Figures S4 and S5). Clade 1 occurs only in the extreme Sands region. This gap exists throughout all time periods and south-eastern tip of the Hajar Mountains and on Masirah Island indicates that these populations have been isolated. The LGM and is predicted to occur only in suitable lowland areas in this model, however, predicts unsuitable climatic conditions at region. Clade 2 is predicted to occur from the south and eastern Masirah Island. Both MESS and MoD pictures depict several Jebel Akhdar to the western Hajars, including the gap between J Zool Syst Evol Res (2016) 54(1), 67--81 © 2015 Blackwell Verlag GmbH 74 DE POUS,MACHADO,METALLINOU, CERVENKA,KRATOCHVIL,PASCHOU,MAZUCH, SMID,SIMO-RIUDALBAS,SANUY and CARRANZA

Fig. 4. Haplotype network of the phased sequences of the nuclear marker cmos. Colours correspond to Fig. 3. Phase probabilities were set as ≥0.7. A list of all the specimens with corresponding sample code, voucher code, country, locality code and GenBank accession numbers is presented in Table 1. these two mountain ranges. Clade 3 has the largest range and is because the authors were not convinced of the generic allocation predicted to occur widespread in the north-western Hajar Moun- of the species, they suggested the combination Trachydactylus tains in Oman and the UAE. The eastern border of this clade is spatalurus. Later, Arnold (1977) argued that the generic charac- predicted to occur along the western ﬂanks and north of the Jebel ters among the Arabian naked-toed geckos were rather arbitrary Akhdar (Figure S5). The potential contact zones as represented and not good indicators of relationships and returned the species by the average probability of the presence of multiple clades tentatively to the genus Bunopus, a taxonomic arrangement fol- were identiﬁed at the margins of the spatial occurrence of the lowed by all subsequent authors (Uetz 2015). Based on our clades as given by the phylogenetic tree using a single threshold molecular data, to resolve the polyphyly of ‘Bunopus’, we resur- (Fig. 5). The potential contact zones with the highest probabili- rect the genus Trachydactylus Haas and Battersby, 1959 (type ties are located in the centre of the Jebel Akhdar (clades 2 and species Trachydactylus jolensis Haas and Battersby, 1959 by 3) as well as in the Eastern Hajars (clades 1 and 2). original designation; considered a junior subjective synonym of Trachydactylus spatalurus (Anderson, 1901) by Arnold 1977) for the clade formed by the two subspecies of Trachydactylus Taxonomic account spatalurus: Trachydactylus spatalurus spatalurus (Anderson, As demonstrated by the results of the multilocus molecular phy- 1901) and Trachydactylus spatalurus hajarensis (Arnold 1980). logenetic analyses and the topological test carried out in this As already stated by Arnold (1980) in the original description study, the genus ‘Bunopus’ is not monophyletic. Bunopus tuber- of Trachydactylus spatalurus hajarensis, the two subspecies of culatus Blanford, 1874, is the type species of the genus Bunopus Trachydactylus spatalurus present some differences at the mor- Blanford, 1874, and therefore, the clade including B. tubercula- phological level. Trachydactylus s. hajarensis is differentiated tus and B. blanfordii should retain the generic name (and would from Trachydactylus s. spatalurus by its smaller size (up to also most probably include Bunopus crassicauda Nikolsky, 50 mm of SVL compared with up to 67 mm in T. s. spatalurus); 1907; Cervenka et al. 2008) and the clade formed by the two the presence in the dorsum of neck and body of about eight (at members of the species ‘Bunopus’ spatalurus should be assigned mid-body) and six (between the hind legs) longitudinal rows of a different generic name. Haas and Battersby (1959), using dis- irregular, enlarged scales with a strong medial keel that increases tinct morphological characters described a new genus and species in height to the posterior border, compared with unkeeled or fee- Trachydactylus jolensis Haas and Battersby 1959. Trachydactylus bly keeled dorsal scales in T. s. spatalurus (Fig. 6a–d); strongly jolensis was found to be morphologically similar to Bunopus keeled scales also on the dorsal part of the hind limbs, compared spatalurus Anderson, 1901 by Leviton and Anderson (1967), but with unkeeled or feebly keeled in T. s. spatalurus (Fig. 6g,j); the J Zool Syst Evol Res (2016) 54(1), 67--81 © 2015 Blackwell Verlag GmbH Taxonomy and biogeography of Bunopus spatalurus 75

Fig. 5. Potential contact zones of three main clades of Trachydactylus hajarensis from the Hajar Mountains in Oman and the UAE. Contact zones are represented by the average probability of presence of multiple clades. The dashed areas show climatic stability as inferred using paleodistribution modelling (Last Glacial Maximum, Mid-Holocene) using the software MAXENT. presence of protruding (convex) scales on top of the head, com- Gamble et al. (2012), with which we share all five nuclear genes pared with flattened scales in T. s. spatalurus (Fig. 6e,h); and of data set 1. The sister taxa relationship between Bunopus and the presence of a pair of enlarged postmental scales, each situ- Crossobamon received high support in all the analyses. However, ated laterally to the posterior section of the mental, compared we observe conflicting levels of support between different analy- with no clearly differentiated chin shields in T. s. spatalurus ses for the phylogenetic position of the other naked-toed gecko (Fig. 6k–n). As a result of the morphological differences, the genera (i.e. Agamura, Cyrtopodion, Tenuidactylus), and we con- high level of genetic differentiation in the 12S mitochondrial sider that it remains unresolved (Gamble et al. 2012; Bauer et al. gene (13.0 Æ 2%), and the results of the phylogenetic and the 2013). As already suggested by Cervenka et al. (2008, 2010) and cmos haplotype network analysis (Figs 2–4 and S2, Table S3), highlighted by Bauer et al. (2013), it seems that there are several we elevate Trachydactylus spatalurus hajarensis to the species cryptic taxa masquerading under B. tuberculatus. According to level Trachydactylus hajarensis (Arnold 1980). our analyses, B. tuberculatus is paraphyletic with respect to B. blanfordii, further supporting this hypothesis. However, until a full assessment of the systematics of B. tuberculatus and Discussion related species (B. blanfordii and B. crassicauda) is carried out, The knowledge on the systematics and phylogenetic relationships including an extensive sampling across its distribution range and of the Palearctic naked-toed geckos has increased in recent years, multilocus phylogenies and morphological analyses, it will not thanks to the use of molecular data (Cervenka et al. 2008, 2010; be possible to reach taxonomic conclusions. Gamble et al. 2012; Bauer et al. 2013). However, despite the The present study shows that Trachydactylus spatalurus and effort in including the maximum number of taxa, neither Trachy- T. hajarensis form two well-supported reciprocally monophyletic dactylus spatalurus nor Trachydactylus hajarensis have been groups that according to our molecular dating diverged during included in any molecular phylogenetic analysis before. As a the Miocene. Regarding T. spatalurus, the scarcity of samples result of that, the taxonomic status of ‘Bunopus’ had not been makes this still a very preliminary study. However, with the evaluated. The phylogenetic relationships recovered in our analy- three samples available, two well-supported groups can be identi- ses of data set 1 (Fig. 2) are very similar to those presented by fied; 1 – locality [30], Yemen; and 2 – localities [28,29], Oman J Zool Syst Evol Res (2016) 54(1), 67--81 © 2015 Blackwell Verlag GmbH 76 DE POUS,MACHADO,METALLINOU, CERVENKA,KRATOCHVIL,PASCHOU,MAZUCH, SMID,SIMO-RIUDALBAS,SANUY and CARRANZA

(a) (c)

(b) (d)

(e) (h) (k)

(l)

(f) (i)

(m)

(g) (j)

(n)

Fig. 6. Pictures of (a) Trachydactylus spatalurus (NMP 75147, male) and (b) Trachydactylus hajarensis (TMHC 2013.10.407). Close up detail of the dorsal scales of (c) Trachydactylus spatalurus (NMP 75147) and (d) Trachydactylus hajarensis (TMHC 2013.10.408). Details of the dorsal side of head (e), lateral side of head (f) and dorsal side of the right hind limb (g) of Trachydactylus spatalurus (NMP 75147); of the dorsal side of head (h), lateral side of head (i) and of the dorsal side of the right hind limb (j) of Trachydactylus hajarensis (TMHC 2013.10.407). Underside of head (gular region showing the arrangement of mental and postmental scales and chin shields) of (k) Trachydactylus spatalurus (NMP 75146, male), (l) Trachydactylus spatalurus (NMP 75147), (m) Trachydactylus hajarensis (TMHC 2013.10.407) and (n) Trachydactylus hajarensis (TMHC 2013.10.409). Data for Trachydactylus spatalurus specimens NMP 75147 and NMP 75146 are presented in Table 1. Trachydactylus hajarensis TMHC 2013.10.407, male, 3 km S. of Al-Hamra, Nizwa, Oman (23.057 57.288, 671 m a.s.l.); TMHC 2013.10.408, male, same data as TMHC 2013.10.407; TMHC 2013.10.409, male, surrounding of Jebel Shams Resort, Oman (23.208 57.200, 1983 m a.s.l.). TMHC: Tomas Mazuch Herpetological Collection, Czech Republic

(Fig. 1) separated by an uncorrected genetic distance in the 12S restricted to the highlands of the Jebel Akhdar, but, as suggested of 2.3–3.2% (see Table S3). According to our estimations, these by phylogenetic analyses using morphological and molecular two lineages split approximately 2.0 Ma according to data set 1 data, it is most probably the result of an independent coloniza- and 2.7 Ma ago according to data set 2 (Figs 2–3). In the south tion from Iran (Arnold and Gardner 1994; Papenfuss et al. of the Arabian Peninsula, there are two main mountain ranges, 2010). Similar patterns are found in the genus Pristurus, with one in western Yemen, in the area called Western Highlands/Ye- P. gallagheri being restricted to the Jebel Akhdar and P. celer- men Highlands, characterized by being geologically closely rimus found in the Jebel Akhdar but also across the Western related to the African side of the Red Sea, and another mountain Hajars and up to the Musandam Peninsula (Arnold 2009), range that extends from central Yemen, in the Central Highlands/ although these two species are not sister taxa (Papenfuss et al. Hadramout uplands, to south Oman (Dhofar Mountains), with 2009; Badiane et al. 2014). The two Hemidactylus endemic to geological affinities to the majority of the Arabian Peninsula the Hajar Mountains and its foothills (H. hajarensis and (Alsinawi and Al Aydrus 1999; Bosworth et al. 2005; As-Saruri H. luqueorum) present almost non-overlapping distribution et al. 2013). Between these two mountain ranges, there is the ranges within the massif. While H. luqueorum is restricted to the Sabatayn basin and more to the north the Sabatayn desert, and highlands of the Jebel Akhdar, its sister species, H. hajarensis,is within each mountain range, a complex mosaic of valleys. These only found in the lowland areas of the Jebel Akhdar and across two areas are important reservoirs of endemism and harbour sig- the Eastern Hajars (Carranza and Arnold 2012). nificant levels of intraspecific genetic variability in geckos (Car- The present work on Trachydactylus hajarensis reveals a high ranza and Arnold 2012; Metallinou et al. 2012, 2015; Sm ıd level of genetic diversity in both mtDNA (12S) and nDNA et al. 2013a). The phylogeographic pattern within T. spatalurus (cmos) genes (Figs 3 and 4), and our molecular dating places its is similar to that reported between the species Uromastyx yeme- intraspecific diversification during the Late Miocene to Pleis- nensis Wilms and Schmitz, 2007 and U. benti (Anderson, 1894), tocene. The three highly divergent clades reported in the present and with similar genetic distances, 2.3–3.3% in the ribosomal study are geographically structured along the Hajar Mountains, mitochondrial gene 16S (Wilms and Schmitz 2007). With an and the total variogram obtained from PHYLIN indicates a pattern increased sampling effort in T. spatalurus, it would be interesting of isolation-by-distance (Figure S4). The cluster variograms sug- to see whether this phylogeographic pattern persists and use this gested by nonlinear least squares show a low fit for clades 1 and and other relevant groups to investigate whether those two 2 which is likely the result of reduced sampling for these clades mountain ranges acted as major independent refugia during cli- that hampers the variogram to depict the spatial dependence for mate shifts, potentially with the Sabatayn basin as a hydrological these lineages. An increased future sampling effort would possi- (advancing of sea water through the basin) or climatic barrier. bly decrease the width of the contact zones and will also clarify By elevating Trachydactylus hajarensis to the specific status, the predicted occurrence of each of the three clades with more we add a new species to the already long and varied list of ende- certainty. The maps of the predicted occurrence (Figure S5) mic reptiles of the Hajar Mountains, which include one snake depict the spatial pattern of the three clades quite well except in (Echis omanensis Babocsay, 2004), one endemic genus of lacer- the under-sampled eastern border areas of clade 2 near the con- tid lizards with two species (Omanosaura jayakari (Boulenger, tact zone with clade 3. Clade 3 occupies the largest area of the 1887) and O. cyanura (Arnold, 1972)), four species of the genus distribution, whereas clade 1 is restricted to the easternmost part Asaccus (A. montanus Gardner, 1994, A. platyrhynchus Arnold of the Hajar Mountains, as well as Masirah Island. We have only and Gardner, 1994, A. gallagheri (Arnold, 1972) and A. caudi- used three main clades in the present PHYLIN sampling scheme, volvulus Arnold and Gardner, 1994), two species of Pristurus while there is substantial genetic variation within these clades (P. gallagheri Arnold, 1986 and P. celerrimus Arnold, 1977) that is not depicted in both the lineage occurrence and contact and three Hemidactylus (H. luqueorum Carranza and Arnold, zone maps. Future work using a broader sampling and a species 2012, H. hajarensis Carranza and Arnold, 2012 and H. endophis delimitation framework (e.g. GMYC; Pons et al. 2006) should Carranza and Arnold, 2012) (Gardner 2013). Moreover, ongoing provide further insights into lineage occupancy while simultane- phylogeographic research in this mountain range suggests that ously narrowing the contact zones between clades. In comparison the levels of genetic variability in some other taxa, including the with many other regions globally, the Hajar Mountains have diurnal gecko Pristurus rupestris Blanford, 1874 and the noctur- remained stable throughout the Late Quaternary (Figure S1) and nal geckos of the genera Ptyodactylus and Asaccus, are much paleodistribution modelling indicates continued suitability higher, with several undescribed species (work in progress). The throughout the distribution range of T. hajarensis. The continued old age and specific situation of the Hajar Mountains, surrounded connection between the ranges of the three clades may have by the sea in the north-west, north and east and by a very large enhanced dispersal (gene flow) throughout the Late Quaternary. arid desert in the south and west (Edgell 2006), have probably The predicted contact zones probably result from the older splits played a crucial role in the origin and maintenance of its unique of the groups when these lineages originated in allopatry, while reptile fauna (Arnold and Gallagher 1977). The Hajar Mountains the maintenance of these contact zones could be the result of have relatively lower vegetation diversity in terms of number of reproductive isolation between the clades. Given the distribution endemic species compared to the Dhofar Mountains (Ghazanfar of the different clades of T. hajarensis, it can be assumed that 1998), but contain some rather isolated highland areas such as they originated through allopatric isolation caused by a combina- the Jebel Akhdar and much lower areas such as the Western tion of past geographical and climatic events during the Miocene Hajars with different climatic conditions, which probably played and Pliocene epoch. A similar pattern of distinctly divergent and still play a very important role in the speciation dynamics of clades across the Hajar Mountains was already found in Hemi- most of the lineages. Among the best examples are the geckos of dactylus hajarensis (Carranza and Arnold 2012), where two the genus Asaccus, with A. platyrhynchus restricted to the Jebel clades occur in allopatry in the eastern and central Hajar Moun- Akhdar massif, A. caudivolvulus distributed across the Western tain ranges. Nevertheless, a case similar to T. hajarensis with the Hajars, and the much smaller-sized A. gallagheri distributed presence of three highly divergent clades across the Hajar Moun- across the whole mountain range and found in sympatry with the tain range has not been reported yet, but preliminary data on other two species (Arnold and Gardner 1994; Papenfuss et al. some other groups (Papenfuss et al. 2010), and especially ongo- 2010). Another endemic Asaccus species, A. montanus,is ing studies using molecular phylogenies for all the reptile ende-

J Zool Syst Evol Res (2016) 54(1), 67--81 © 2015 Blackwell Verlag GmbH 78 DE POUS,MACHADO,METALLINOU, CERVENKA,KRATOCHVIL,PASCHOU,MAZUCH, SMID,SIMO-RIUDALBAS,SANUY and CARRANZA mics of this massif, indicate that this could be a common pattern. be maintained despite increased sampling in extreme north-east- Additional studies using other taxa will help to understand the ern Oman. processes that have shaped the distribution patterns of the taxa of The present study indicates that the combinations of morphol- the Hajar Mountains. ogy, molecular phylogenetics, paleodistribution modelling and Although the mitochondrial and nuclear results presented in phylogeographic interpolation are very powerful tools for use in Figs 3 and 4 and in Table S3 seem to suggest that T. hajarensis taxonomy, biogeography and phylogeography. Future studies in the Hajar Mountains can represent a species complex, more comparing the patterns of the several endemic reptiles of the geo- detailed phylogenetic and morphological analyses will be neces- logically complex Hajar Mountains using an integrative approach sary to investigate this issue. The most divergent clade (clade 1) that combines spatial modelling and genetics will be very includes specimens from the easternmost part of Oman and a valuable to obtain a better understanding of the patterns and pro- population on Masirah Island. These populations seem to be iso- cesses that have shaped their unique diversity and distribution. lated by the Sharqiyah Sands, a sand dune desert with no records of T. hajarensis (Arnold 1980; Gallagher and Arnold 1988; Gardner 2013; pers. observ.) According to our dating Acknowledgements analyses (Fig. 3) and p-distance values (Table S3), the three We wish to thank Felix Amat, Elena Gomez-D ıaz, Raquel Vasconcelos, specimens of clade 1 from Masirah Island (locs. 25–27 in David Hegner, Katerina Kadleckova, Michal Krelina and Pavel Novak for Fig. 1) are differentiated from the rest of specimens from clade assisting in sample collection in the field. Special thanks are due to Saleh 1 (locs. 23–24 in Fig. 1) by a p-distance in the 12S of 0.9%, Al Saadi, Mohammed Al Shariani, Thuraya Al Sariri, Ali Al Kiyumi, having diverged around 0.8 Ma (0.2–2.2, 95% HPD), during the Mohammed Abdullah Al Maharmi and the other members of the Nature Pleistocene. These results, together with the widespread distribu- Conservation Department of the Ministry of Environment and Climate, Sul- tanate of Oman for their help and support and for issuing all the necessary tion range of T. hajarensis within Masirah Island (pers. observ.), permits (Refs: 08/2005; 16/2008; 38/2010; 12/2011; 13/2013; 21/2013). suggest that the population from Masirah Island may have been This work was supported by the project ‘Field study for the conservation of established by natural colonization instead of being the result of reptiles in Oman’ funded by the Ministry of Environment and Climate a human-mediated introduction (cf. to lizards of the genus Cha- Affairs, Oman (Ref: 22412027), and grant CGL2012-36970 from the Min- maeleo; Gardner 2013). The LGM model shows unsuitable con- isterio de Economıa y Competitividad, Spain (cofunded by FEDER). PdP ditions on Masirah Islands, and it is therefore hypothesized that is funded by a FI-DGR grant from the Generalitat de Catalunya, Spain this population persisted in suitable microrefugia and subse- (2014FI_B2 00197). The work of JS was financially supported by Ministry quently recolonized most parts of the island. The relative coarse of Culture of the Czech Republic (DKRVO 2015/15, National Museum, 00023272). LM is funded by Fundacß~ao para a Ciência e Tecnologia (FCT) resolution used for SDM has probably resulted in the unsuitabil- fi ity of the island, but using finer grained GCM would probably through a PhD grant (SFRH/BD/89820/2012) nanced by Programa Opera- cional Potencial Humano (POPH) – Quadro de Referência Estrategico show that heterogeneity in the mountainous areas on Masirah Nacional (QREN) from the European Social Fund and Portuguese Islands provided temporary microrefugia for T. hajarensis and Ministerio da Educacß~ao e Ciência. MSR is funded by a FPI grant from the other species (see Hannah et al. 2014). The predictive power of Ministerio de Economıa y Competitividad, Spain (BES-2013-064248). paleodistribution models has been recently criticized when applying such models to species that violate the basic assump- tion of the environment as main driver of their distribution pat- References terns (Tonini et al. 2013). Trachydactylus hajarensis is a Akaike H (1973) Information theory and an extension of the maximum widespread generalist species that is largely dependent on the likelihood principle. In: Petrov BN, Csaki F (eds), Information Theory presence of rocky substrates in areas with sufficient landscape and an Extension of the Maximum Likelihood Principle. Akademiai heterogeneity and precipitation. The climate in most deserts is Kiado, Budapest, pp 267–281. relatively homogeneous compared to tropic or temperate regions, Alsinawi SA, Al Aydrus A (1999) Seismicity of Yemen. Obadi Studies and most reptile species in Oman have mainly specialized in and Publishing Center, Sana’a, Republic of Yemen. microhabitat use such as sandy versus rocky plains (Metallinou Anderson J (1896) A Contribution to the Herpetology of Arabia, With a et al. 2012, 2015). In the lowland regions of Oman, T. hajaren- Preliminary List of the Reptiles and Batrachians of Egypt. R.H. Porter, London, 124 p. sis is directly replaced by the competing B. tuberculatus and the Arnold EN (1972) Lizards with northern affinities from the mountains of absence of the latter species on Masirah Island could well Oman. Zool Med 47:111–128. explain the continued persistence of populations of the former. Arnold EN (1977) Little-known geckos (Reptilia: Gekkonidae) from Nevertheless, more sampling in north-east Oman will be neces- Arabia with descriptions of two new species from the Sultanate of sary to assess the actual level of genetic differentiation between Oman. J Oman Stud Spec Rep 1:81–110. T. hajarensis from this area and from Masirah Island. Other Arnold EN (1980) The scientific results of the Oman flora and fauna taxa, such as for instance the amphibian Duttaphrynus dhufaren- survey 1977 (Dhofar). The reptiles and amphibians of Dhofar, sis (Parker, 1931), and the reptiles, such as Telescopus dhara southern Arabia. J Oman Stud Spec Rep 2:273–332. dhara (Forskal, 1775), Spalerosophis diadema cliffordii (Sch- Arnold EN (1986) A key and annotated checklist to the lizards and – legel, 1837) and Platyceps rhodorachis, seem to present a simi- amphisbaenians of Arabia. Fauna Saudi Arabia 8:385 435. Arnold EN (2009) Relationships, evolution and biogeography of lar disjunct distribution to T. hajarensis with a hiatus between Semaphore geckos, Pristurus (Squamata, Sphaerodactylidae) based on Masirah Island and the easternmost Hajar Mountains (Gardner morphology. Zootaxa 2060:1–21. 2013). Geomorphological studies and dating estimates of the Arnold EN, Gallagher MD (1977) The scientific results of the Oman presence of endemic sand dune specialist taxa such as Sten- flora and fauna survey 1975. Reptiles and amphibians from the odactylus sharqiyahensis Metallinou and Carranza, 2013 suggest mountains of northern Oman with special reference to the Jebel that the Sharqiyah Sands have been in place for a relatively long Akhdar region. J Oman Stud Spec Rep 1:59–80. period, which could date back to the Middle Miocene to Plio- Arnold EN, Gardner AS (1994) A review of the Middle Eastern leaf-toed Pleistocene (Radies et al. 2004; Metallinou and Carranza 2013). geckoes (Gekkonidae: Asaccus) with descriptions of two new species – If the connection between the populations of T. hajarensis from from Oman. Fauna Saudi Arabia 14:424 441. As-Saruri MA, Sorkhabi R, Baraba R (2013) Sedimentary basins of the Hajar Mountains and Masirah Island had been cut off at that Yemen: their tectonic development and lithostratigraphic cover. In: Al time, we would expect that the level of genetic variability would

Hosani K, Roure G, Ellison R, Lokier S (eds), Lithosphere Dynamics Cox SC, Carranza S, Brown RP (2010) Divergence times and and Sedimentary Basins: The Arabian Plate and Analogues. Springer, colonization of the Canary Islands by Gallotia lizards. Mol Phylogenet Berlin Heidelberg, pp 361–373. Evol 56:747–757. Babocsay G (2004) A new species of saw-scaled viper of the Echis Darriba D, Taboada GL, Doallo R, Posada D (2012) jModelTest 2: more coloratus complex (Ophidia: Viperidae) from Oman, Eastern Arabia. models, new heuristics and parallel computing. Nat Methods 9:772. Syst Biodivers 1:503–514. Drummond A, Rambaut A (2007) BEAST: Bayesian evolutionary Badiane A, Garcia-Porta J, Cervenka J, Kratochvıl L, Sindaco R, Robinson analysis by sampling trees. BMC Evol Biol 7:214. MD, Morales H, Mazuch T, Price T, Amat F, Shobrak MY, Wilms T, Edgell HS (2006) Arabian Deserts: Nature, Origin and Evolution. Simo-Riudalbas M, Ahmadzadeh F, Papenfuss TJ, Cluchier A, Viglione Springer, Dordrecht, 592 p. J, Carranza S (2014) Phylogenetic relationships of Semaphore geckos Elith J, Kearney M, Phillips S (2010) The art of modelling range shifting (Squamata: Sphaerodactylidae: Pristurus) with an assessment of the species. Methods Ecol Evol 1:330–342. taxonomy of Pristurus rupestris. Zootaxa 3835:033–058. Felsenstein J (1985) Confidence limits on phylogenies: an approach using Barata M, Perera A, Martınez-Freirıa F, Harris DJ (2012) Cryptic the bootstrap. Evolution 39:783–791. diversity within the Moroccan endemic day geckos Quedenfeldtia Ficetola GF, Bonardi A, Sindaco R, Padoa-Schioppa E (2013) Estimating (Squamata: Gekkonidae): a multidisciplinary approach using genetic, patterns of reptile biodiversity in remote regions. J Biogeogr 40:1202– morphological and ecological data. Biol J Linn Soc 106:828–850. 1211. Bauer AM, de Silva A, Greenbaum E, Jackman T (2007) A new species Flot JF (2010) SeqPHASE: a web tool for interconverting PHASE input/ of day gecko from high elevation in Sri Lanka, with a preliminary output files and FASTA sequence alignments. Mol Ecol Res 10:162– phylogeny of Sri Lankan Cnemaspis (Reptilia, Squamata, 166. Gekkonidae). Zoosyst Evol 83:22–32. Gallagher MD, Arnold EN (1988) Reptiles and amphibians from the Bauer AM, Masroor R, Titus-Mcquillan J, Heinicke MP, Daza JD, Wahiba Sands, Oman. J Oman Stud Spec Rep 3:405–413. Jackman TD (2013) A preliminary phylogeny of the Palearctic naked- Gamble T, Bauer AM, Greenbaum E, Jackman TR (2008) Evidence of toed geckos (Reptilia: Squamata: Gekkonidae) with taxonomic Gondwanan vicariance in an ancient clade of geckos. J Biogeogr implications. Zootaxa 3599:301–324. 35:88–104. Bosworth W, Huchon P, McClay K (2005) The Red Sea and Gulf of Gamble T, Greenbaum E, Jackman TR, Russell AP, Bauer AM (2012) Aden Basins. J Afr Earth Sci 43:334–378. Repeated origin and loss of adhesive toepads in geckos. PLoS ONE 7: Brown JL (2014) SDMtoolbox: a python-based GIS toolkit for landscape e39429. genetic, biogeographic and species distribution model analyses. Gardner AS (2013) The Amphibians and Reptiles of Oman and the Methods Ecol Evol 5:694–700. UAE. Edition Chimaira, Frankfurt am Main, 480 p. Brown RP, Pestano J (1998) Phylogeography of skinks (Chalcides)in Gasperetti J (1988) Snakes of Arabia. Fauna Saudi Arabia 9:169–450. the Canary Islands inferred from mitochondrial DNA sequences. Mol Ghazanfar SA (1998) Status of the flora and plant conservation in the Ecol 7:1183–1191. sultanate of Oman. Biol Cons 85:287–295. Brown RP, Yang Z (2010) Bayesian dating of shallow phylogenies with Groth JG, Barrowclough GF (1999) Basal divergences in birds and the a relaxed clock. Syst Biol 59:119–131. phylogenetic utility of the nuclear RAG-1 gene. Mol Phylogenet Evol Brown RP, Terrasa B, Perez-Mellado V, Castro JA, Hoskisson PA, 12:115–123. Picornell A, Ramon MM (2008) Bayesian estimation of post- Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to Messinian divergence times in Balearic Island lizards. Mol Phylogenet estimate large phylogenies by maximum likelihood. Syst Biol 52:696–704. Evol 48:350–358. Haas G (1957) Some amphibians and reptiles from Arabia. Proc Calif Busais SM, Joger U (2011) Three new species and one new subspecies Acad Sci 29:47–86. of Hemidactylus Oken, 1817 from Yemen (Squamata, Gekkonidae). Haas G, Battersby JC (1959) Amphibians and reptiles from Arabia. Vertebr Zool 61:267–280. Copeia 1959:196–202. Carnaval AC, Hickerson MJ, Haddad CF, Rodrigues MT, Moritz C Hannah L, Flint L, Syphard AD, Moritz MA, Buckley LB, McCullough (2009) Stability predicts genetic diversity in the Brazilian Atlantic IM (2014) Fine-grain modeling of species’ response to climate change: forest hotspot. Science 323:785–789. holdouts, stepping-stones, and microrefugia. Trends Ecol Evol 29:390– Carranza S, Arnold EN (2012) A review of the geckos of the genus 397. Hemidactylus (Squamata: Gekkonidae) from Oman based on Harrigan RJ, Mazza ME, Sorenson MD (2008) Computation vs. cloning: morphology, mitochondrial and nuclear data, with descriptions of eight evaluation of two methods for haplotype determination. Mol Ecol Res new species. Zootaxa 3378:1–95. 8:1239–1248. Carranza S, Arnold EN, Mateo J, Lopez-Jurado L (2000) Long-distance Hawlitschek O, Glaw F (2013) The complex colonization history of colonization and radiation in gekkonid lizards, Tarentola (Reptilia: nocturnal geckos (Paroedura) in the Comoros Archipelago. Zool Scr Gekkonidae), revealed by mitochondrial DNA sequences. P Roy Soc 42:135–150. B-Biol Sci 267:637–649. Heled J, Drummond AJ (2010) Bayesian inference of species trees from Carranza S, Arnold EN, Mateo JA, Geniez P (2002) Relationships and multilocus data. Mol Biol Evol 27:570–580. evolution of the North African geckos, Geckonia and Tarentola Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very (Reptilia: Gekkonidae), based on mitochondrial and nuclear DNA high resolution interpolated climate surfaces for global land areas. Int J sequences. Mol Phylogenet Evol 23:244–256. Climatol 25:1965–1978. Carranza S, Arnold EN, Geniez P, Roca J, Mateo JA (2008) Radiation, Huelsenbeck JP, Rannala B (2004) Frequentist properties of Bayesian multiple dispersal and parallelism in the skinks, Chalcides and posterior probabilities of phylogenetic trees under simple and complex Sphenops (Squamata: Scincidae), with comments on Scincus and substitution models. Syst Biol 53:904–913. Scincopus and the age of the Sahara Desert. Mol Phylogenet Evol Huelsenbeck JP, Bollback JP, Levine AM (2002) Bayesian method for 46:1071–1094. inferring the root of a phylogenetic tree. Syst Biol 51:32–43. Castresana J (2000) Selection of conserved blocks from multiple Jimenez-Valverde A, Lobo JM (2007) Threshold criteria for conversion alignments for their use in phylogenetic analysis. Mol Biol Evol of probability of species presence to either-or presence-absence. Acta 17:540–552. Oecol 31:361–369. Cervenka J, Kratochvıl L, Frynta D (2008) Phylogeny and taxonomy of Katoh K, Standley DM (2013) MAFFT multiple sequence alignment the Middle Eastern geckos of the genus Cyrtopodion and their selected software version 7: improvements in performance and usability. Mol relatives. Zootaxa 1931:25–36. Biol Evol 30:772–780. Cervenka J, Frynta D, Kratochvıl L (2010) Phylogenetic relationships of Lanfear R, Calcott B, Ho SY, Guindon S (2012) PartitionFinder: the gecko genus Carinatogecko (Reptilia: Gekkonidae). Zootaxa combined selection of partitioning schemes and substitution models for 2636:59–64. phylogenetic analyses. Mol Biol Evol 29:1695–1701. Clement M, Posada D, Crandall KA (2000) TCS: a computer program to Leviton AE, Anderson SC (1967) Survey of the reptiles of the estimate gene genealogies. Mol Ecol 9:1657–1660. sheikhdom of Abu Dhabi, Arabian Peninsula. Part II. Systematic

account of reptiles made in the sheikhdom of Abu Dhabi by John Shimodaira H, Hasegawa M (1999) Multiple comparisons of log- Gasperetti. Proc Calif Acad Sci 35:157–192. likelihoods with applications to phylogenetic inference. Mol Biol Evol Liu C, Berry PM, Dawson TP, Pearson RG (2005) Selecting thresholds 16:1114–1116. of occurrence in the prediction of species distributions. Ecography Shimodaira H, Hasegawa M (2001) CONSEL: for assessing the conﬁdence 28:385–393. of phylogenetic tree selection. Bioinformatics 17:1246–1247. Merow C, Smith MJ, Silander JA (2013) A practical guide to MaxEnt Silvestro D, Michalak I (2012) raxmlGUI: a graphical front-end for for modeling species’ distributions: what it does, and why inputs and RAxML. Org Divers Evol 12:335–337. settings matter. Ecography 36:1058–1069. Sindaco R, Jeremcenko VK (2008) The Reptiles of the Western Metallinou M, Carranza S (2013) New species of Stenodactylus Palearctic. Edizioni Belvedere, Latina, Italy, 579 p. (Squamata: Gekkonidae) from the Sharqiyah Sands in northeastern Sindaco R, Metallinou M, Pupin F, Fasola M, Carranza S (2012) Oman. Zootaxa 3745:449–468. Forgotten in the ocean: systematics, biogeography and evolution of Metallinou M, Arnold EN, Crochet PA, Geniez P, Brito JC, Lymberakis the Trachylepis skinks of the Socotra Archipelago. Zool Scr 41: P, Baha El Din S, Sindaco R, Robinson M, Carranza S (2012) 346–362. Conquering the Sahara and Arabian deserts: systematics and Sm ıd J, Carranza S, Kratochvıl L, Gvozdık V, Nasher AK, Moravec J biogeography of Stenodactylus geckos (Reptilia: Gekkonidae). BMC (2013a) Out of Arabia: a complex biogeographic history of multiple Evol Biol 12:258. vicariance and dispersal events in the gecko genus Hemidactylus Metallinou M, Cervenka J, Crochet PA, Kratochvıl L, Wilms T, Geniez (Reptilia: Gekkonidae). PLoS ONE 8:e64018. P, Shobrak MY, Brito JC, Carranza S (2015) Species on the rocks: Sm ıd J, Moravec J, Kratochvıl L, Gvozdık V, Nasher AK, Busais SM, systematics and biogeography of the rock-dwelling Ptyodactylus Wilms T, Shobrak MY, Carranza S (2013b) Two newly recognized geckos (Squamata: Phyllodactylidae) in North Africa and Arabia. Mol species of Hemidactylus (Squamata, Gekkonidae) from the Arabian Phylogenet Evol 85:208–220. Peninsula and Sinai, Egypt. ZooKeys 355:79. Mila B, Surget-Groba Y, Heulin B, Gosa A, Fitze PS (2013) Multilocus Sm ıd J, Moravec J, Kratochvıl L, Nasher AK, Mazuch T, GvozdıkV, phylogeography of the common lizard Zootoca vivipara at the Ibero- Carranza S (2015) Multilocus phylogeny and taxonomic revision of Pyrenean suture zone reveals lowland barriers and high-elevation the Hemidactylus robustus species group (Reptilia, Gekkonidae) with introgression. BMC Evol Biol 13:192. descriptions of three new species from Yemen and Ethiopia. Syst Muscarella R, Galante PJ, Soley-Guardia M, Boria RA, Kass JM, Uriarte M, Biodivers 14:346–368. Anderson RP (2014) ENMeval: an R package for conducting spatially Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based independent evaluations and estimating optimal model complexity for phylogenetic analyses with thousands of taxa and mixed models. Maxent ecological niche models. Methods Ecol Evol 5:1198–1205. Bioinformatics 22:2688–2690. Papenfuss TJ, Jackman T, Bauer A, Stuart BL, Robinson MD, Parham JF Stephens M, Smith NJ, Donnelly P (2001) A new statistical method for (2009) Phylogenetic relationships among species in the Sphaerodactylid haplotype reconstruction from population data. Am J Hum Genet lizard genus Pristurus. Proc Calif Acad Sci 60:675–681. 68:978–989. Papenfuss TJ, Jackman TR, Bauer AM, Stuart BL, Robinson MD, Talavera G, Castresana J (2007) Improvement of phylogenies after Parham JF (2010) Phylogenetic relationships among species of removing divergent and ambiguously aligned blocks from protein Southwest Asian Leaf-toed geckos (Asaccus). Proc Calif Acad Sci sequence alignments. Syst Biol 56:564–577. 61:587–596. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) Phillips SJ, Anderson RP, Schapire RE (2006) Maximum entropy MEGA5: molecular evolutionary genetics analysis using maximum modeling of species geographic distributions. Ecol Model 190:231–259. likelihood, evolutionary distance, and maximum parsimony methods. Pianka ER (1973) The structure of lizard communities. Annu Rev Ecol Mol Biol Evol 28:2731–2739. Evol Syst 4:53–74. Tarroso P, Velo-Anton G, Carvalho SB (2015) PHYLIN: an r package Pianka ER (1986) Ecology and Natural History of Desert Lizards: for phylogeographic interpolation. Mol Ecol Res 15:349–357. Analyses of the Ecological Niche and Community Structure. Princeton Tonini JFR, Costa LP, Carnaval AC (2013) Phylogeographic structure is University Press, Princeton, 222 p. strong in the Atlantic Forest; predictive power of correlative Pons J, Barraclough TG, Gomez-Zurita J, Cardoso A, Duran DP, Hazell paleodistribution models, not always. J Zool Syst Evol Res 51:114– S, Kamoun S, Sumlin WD, Vogler AP (2006) Sequence-based species 121. delimitation for the DNA taxonomy of undescribed insects. Syst Biol Uetz P (2015) The Reptile Database. Available from http://www.reptile- 55:595–609. database.org. Last accessed June 22, 2015 de Queiroz K (1998) The general lineage concept of species, species VanDerWal J, Shoo LP, Graham C, Williams SE (2009) Selecting criteria, and the process of speciation: a conceptual uniﬁcation and pseudo-absence data for presence only distribution modelling, how far terminological recommendations. In: Howard DJ, Berlocher SH (eds), should we stray from what we know? Ecol Model 220:589–594. Endless Forms: Species and Speciation. Oxford University Press, Vasconcelos R, Carranza S (2014) Systematics and biogeography of Oxford, pp 57–75. Hemidactylus homoeolepis Blanford, 1881 (Squamata: Gekkonidae), de Queiroz K (2007) Species concepts and species delimitation. Syst Biol with the description of a new species from Arabia. Zootaxa 3835:501– 56:879–886. 527. Rabosky DL, Cowan MA, Talaba AL, Lovette IJ (2011) Species Waltari E, Hijmans RJ, Peterson AT, Nyari AS, Perkins SL, Guralnick interactions mediate phylogenetic community structure in a hyperdiverse RP (2007) Locating Pleistocene refugia: comparing phylogeographic lizard assemblage from arid Australia. Am Nat 178:579–595. and ecological niche model predictions. PLoS ONE 2:e563. Radies D, Preusser F, Matter A, Mange M (2004) Eustatic and climatic Warren DL, Seifert SN (2011) Ecological niche modeling in Maxent: the controls on the development of the Wahiba Sand sea, Sultanate of importance of model complexity and the performance of model Oman. Sedimentology 51:1359–1385. selection criteria. Ecol Appl 21:335–342. Rambaut A, Drummond A (2007) Tracer v1.4. Available from: http:// Warren DL, Glor RE, Turelli M (2010) ENMTools: a toolbox for beast.bio.ed.ac.uk/Tracer comparative studies of environmental niche models. Ecography Sambrook J, Fritsch EF, Maniatis T (1989) Molecular Cloning: A 33:607–611. Laboratory Manual. Cold Spring Harbour Press, New York, 545 p. Wilcox TP, Zwickl DJ, Heath TA, Hillis DM (2002) Phylogenetic Sch€atti B, Desvoignes A (1999) The Herpetofauna of Southern Yemen relationships of the dwarf boas and a comparison of Bayesian and and the Sokotra Archipelago. Gilbert-E. Huguet, Geneve, 179 p. bootstrap measures of phylogenetic support. Mol Phylogenet Evol Sch€atti B, Gasperetti J (1994) A contribution to the herpetofauna of 25:361–371. Southwest Arabia. Fauna Saudi Arabia 14:348–423. Wilms T, Schmitz A (2007) A new polytypic species of the genus Shimodaira H (2002) An approximately unbiased test of phylogenetic Uromastyx Merrem 1820 (Reptilia: Squamata: Agamidae: tree selection. Syst Biol 51:492–508. Leiolepidinae) from southwestern Arabia. Zootaxa 1394:1–23.

Supporting Information wise samples within each distance class is represented with different circle size. Additional Supporting Information may be found in the online Figure S5. Predicted occurrence of the three main clades of version of this article: Trachydactylus hajarensis using phylogeographic interpolation Figure S1. Potential species distribution models of Trachy- using the PHYLIN R package. dactylus hajarensis. A) The available distribution records used Table S1. Primers used in this study. for SDM, B) present, C), Mid-Holocene and D) Last Glacial Table S2. Distribution database of Trachydactylus hajarensis Maximum. All models are above the MTSPS threshold. with latitude, longitude, as used for species distribution mod- Figure S2. A) Bayesian tree inferred using BEAST on data elling in the present study. set 1 (12S, cmos, rag1, rag2, acm4, and pdc genes); B) species Table S3. 12S uncorrected genetic distances (p-distances; * tree inferred with BEAST using data set 1. The two Hemidacty- complete deletion) between all the specimens of Trachydactylus lus used to root the tree are not shown. Empty circles indicate spatalurus and Trachydactylus hajarensis included in the present > pp 0.95. Taxon names correspond to changes proposed in this study. Colours are according to the different clades highlighted paper. in Figure 3. In red: Trachydactylus spatalurus; dark blue, clade Figure S3. MESS (A and C) and MoD (B and D) pictures of 1ofTrachydactylus hajarensis; blue, clade 2 of Trachydactylus the projected SDMs for the (A-B) Mid-Holocene and (C-D) Last hajarensis; light blue, clade 3 of Trachydactylus hajarensis. Glacial Maximum. Genetic distances are given in % and are shown below the diago- Figure S4. Total (A) and cluster variograms (B clade 1, C nal. The Standard Errors are shown above the diagonal. clade 2 and D clade 3) with ﬁtted models. The number of pair-

Taxonomy and biogeography of Bunopus spatalurus (Reptilia; Gekkonidae) from the Arabian Peninsula

Philip de Pous1,2, Luis Machado1,3,4, Margarita Metallinou5, Jan Červenka6, Lukáš Kratochvíl6, Nefeli Paschou1, Tomáš Mazuch7, Jiří Šmíd8, Marc Simó-Riudalbas1, Delfi Sanuy2, Salvador 1* Carranza

1Institute of Evolutionary Biology (CSIC-Universitat Pompeu Fabra), Passeig Marítim de la Barceloneta 37-49, 08003 Barcelona, Spain. 2Faculty of Life Sciences and Engineering, Departament Producció Animal (Fauna Silvestre), Universitat de Lleida, Av. Rovira Roura 191, 25198, Lleida, Spain. 3CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, Universidade do Porto, InBio Laboratório Associado, Campus Agrário de Vairão, 4485-661 Vairão, Portugal. 4Departamento de Biologia, Faculdade de Ciencias da Universidade do Porto, Rua do Campo Alegre, 4169-007 Porto, Portugal. 5Department of Biology, Villanova University, 800 Lancaster Avenue, Villanova, PA 19085, USA. 6Department of Ecology, Faculty of Science, Charles University in Prague, Viničná 7, CZ-128 44 Praha 2, Czech Republic. 7Dříteč 65, 53305, Czech Republic. 8Department of Zoology, National Museum, Prague, Czech Republic.

*Corresponding author. E-mail: [email protected]

Figure S1. Potential species distribution models of Trachydactylus hajarensis. A) The available distribution records used for SDM, B) present, C), Mid-Holocene and D) Last Glacial Maximum. All models are above the MTSPS threshold.

Figure S2. A) Bayesian tree inferred using BEAST on dataset 1 (12S, cmos, rag1, rag2, acm4, and pdc genes); B) species tree inferred with *BEAST using dataset 1. The two Hemidactylus used to root the tree are not shown. Empty circles indicate pp > 0.95. Taxon names correspond to changes proposed in this paper.

Figure S3. MESS (A and C) and MoD (B and D) pictures of the projected SDMs for the (A-B) Mid-Holocene and (C-D) Last Glacial Maximum.

Figure S4. Total (A) and cluster variograms (B clade 1, C clade 2 and D clade 3) with fitted models. The number of pairwise samples within each distance class is represented with different circle size.

Figure S5. Predicted occurrence of the three main clades of Trachydactylus hajarensis using phylogeographic interpolation using the PHYLIN R package.

Table S1. Primers used in this study. For and Rev refer to forward and reverse, respectively.

Gene Primer Orientation Reference Sequence (5' - 3') 12S 12SaGekko Forward Metallinou et al. (2015) CAAACTAGGATTAGATACCCTACTATGC 12S 12SbGekko Reverse Metallinou et al. (2015) GAGGGTGACGGGCGGTGTGTAC PDC PHOF2 Forward Bauer et al. (2007) AGATGAGCATGCAGGAGTATGA PDC PHOR1 Reverse Bauer et al. (2007) TCCACATCCACAGCAAAAAACTCCT c-mos FU-F Forward Gamble et al. (2008) TTTGGTTCKGTCTACAAGGCTAC c-mos FU-R Reverse Gamble et al. (2008) AGGGAACATCCAAAGTCTCCAA RAG1 F700 Forward Bauer et al. (2007) GGAGACATGGACACAATCCATCCTAC RAG1 R700 Reverse Bauer et al. (2007) TTTGTACTGAGATGGATCTTTTTGCA RAG1 R13 Forward Groth and Barrowclough (1999) TCTGAATGGAAATTCAAGCTGTT RAG1 R18 Reverse Groth and Barrowclough (1999) GATGCTGCCTCGGTCGGCCACCTTT RAG2 PY1-F Forward Gamble et al. (2008) CCCTGAGTTTGGATGCTGTACTT RAG2 PY1-R Reverse Gamble et al. (2008) AACTGCCTRTTGTCCCCTGGTAT ACM4 Int-F Forward Gamble et al. (2008) TTTYCTGAAGAGCCCTCTGGTC ACM4 Int-R Reverse Gamble et al. (2008) CAAATTTCCTGGCAACATTRGC

Table S2. Distribution database of Trachydactylus hajarensis with latitude, longitude, as used for species distribution modelling in the present study.

Species,Lat,Long Trachydactylus_hajarensis,3280202.7761,-7532487.3804 Trachydactylus_hajarensis,3231209.5397,-7554738.8005 Trachydactylus_hajarensis,3064566.1726,-7565693.468 Trachydactylus_hajarensis,3040367.0931,-7581372.7552 Trachydactylus_hajarensis,3257284.7792,-7779619.587 Trachydactylus_hajarensis,3278523.733,-7767290.5958 Trachydactylus_hajarensis,3307176.1835,-7591839.6166 Trachydactylus_hajarensis,3301442.7013,-7742270.7396 Trachydactylus_hajarensis,3308718.4751,-7661615.6955 Trachydactylus_hajarensis,3272691.5366,-7596113.5738 Trachydactylus_hajarensis,3270829.0658,-7804580.0293 Trachydactylus_hajarensis,3302839.2581,-7921388.1982 Trachydactylus_hajarensis,3416116.4791,-7976965.6299 Trachydactylus_hajarensis,3378491.7183,-7937809.956 Trachydactylus_hajarensis,3517020.7973,-7983689.999 Trachydactylus_hajarensis,3458153.1504,-7991511.9652 Trachydactylus_hajarensis,3305635.8531,-7775345.3866 Trachydactylus_hajarensis,3262733.3723,-7857223.229 Trachydactylus_hajarensis,3439570.3321,-8003551.6101 Trachydactylus_hajarensis,3497725.5047,-7997667.1796 Trachydactylus_hajarensis,3433922.4495,-7979941.786 Trachydactylus_hajarensis,3275426.4664,-7823853.9284 Trachydactylus_hajarensis,3338514.0841,-7682654.7623 Trachydactylus_hajarensis,3477600.3992,-7990827.8911 Trachydactylus_hajarensis,3291267.2681,-7609297.6795 Trachydactylus_hajarensis,3320295.9883,-8007284.3403 Trachydactylus_hajarensis,3411615.7907,-8024236.6987 Trachydactylus_hajarensis,3452097.2425,-8012174.5411 Trachydactylus_hajarensis,3323157.9055,-7978542.6578 Trachydactylus_hajarensis,3346418.2113,-7981669.7205 Trachydactylus_hajarensis,3364237.4847,-7985973.6116 Trachydactylus_hajarensis,3493809.9959,-8014532.322 Trachydactylus_hajarensis,3436018.0471,-7998106.4727 Trachydactylus_hajarensis,3489218.9796,-7995232.5976 Trachydactylus_hajarensis,3439245.2708,-7974816.003 Trachydactylus_hajarensis,3385014.7954,-7954337.927 Trachydactylus_hajarensis,3466674.377,-7973499.8413 Trachydactylus_hajarensis,3517020.7973,-7983689.999 Trachydactylus_hajarensis,3302839.2581,-7921388.1982 Trachydactylus_hajarensis,3289098.8539,-7899106.519 Trachydactylus_hajarensis,3262733.3723,-7857223.229 Trachydactylus_hajarensis,3275225.205,-7824044.3434 Trachydactylus_hajarensis,3296554.885,-7814819.119 Trachydactylus_hajarensis,3270829.0658,-7804580.0293 Trachydactylus_hajarensis,3196399.1115,-7786227.8985 Trachydactylus_hajarensis,3257284.7792,-7779619.587 Trachydactylus_hajarensis,3295202.291,-7786517.7345 Trachydactylus_hajarensis,3275620.9311,-7774498.5987 Trachydactylus_hajarensis,3278961.3423,-7751956.8365 Trachydactylus_hajarensis,3320673.8755,-7757579.3045 Trachydactylus_hajarensis,3298636.0963,-7744953.2282 Trachydactylus_hajarensis,3338471.1687,-7725057.3169 Trachydactylus_hajarensis,3344974.3221,-7684623.2602 Trachydactylus_hajarensis,3315131.9735,-7663388.6481 Trachydactylus_hajarensis,3031202.6721,-7591006.1892 Trachydactylus_hajarensis,3056671.0326,-7585256.6537 Trachydactylus_hajarensis,3081693.6093,-7573436.2121 Trachydactylus_hajarensis,3064566.1726,-7565693.468 Trachydactylus_hajarensis,3257476.7444,-7603676.258 Trachydactylus_hajarensis,3291267.2681,-7609297.6795 Trachydactylus_hajarensis,3316993.3582,-7613988.135 Trachydactylus_hajarensis,3308929.8422,-7596501.7468 Trachydactylus_hajarensis,3231209.5397,-7554738.8005 Trachydactylus_hajarensis,3280202.7761,-7532487.3804 Trachydactylus_hajarensis,3256781.5692,-7507920.8667

Table S3. 12S uncorrected genetic distances (p-distances; complete deletion) between all the specimens of Trachydactylus spatalurus and Trachydactylus hajarensis included in the present study. Colors are according to the different clades highlighted in Figure 3. In red: Trachydactylus spatalurus; dark blue, clade 1 of Trachydactylus hajarensis; blue, clade 2 of Trachydactylus hajarensis; light blue, clade 3 of Trachydactylus hajarensis. Genetic distances are given in % and are shown below the diagonal. The Standard Errors are shown above the diagonal.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 1 TMHC404 0.5 0.8 0.8 1.6 1.6 1.6 1.6 1.6 1.8 1.8 1.8 1.8 1.8 1.7 1.8 1.8 1.8 1.8 1.8 1.8 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 2 TMHC405 0.9 0.9 0.9 1.7 1.7 1.7 1.7 1.7 1.8 1.8 1.8 1.8 1.8 1.7 1.8 1.8 1.8 1.8 1.8 1.8 1.8 1.8 1.8 1.8 1.7 1.7 1.7 1.7 1.7 1.7 1.7 3 SPAT01 2.3 3.2 0.0 1.7 1.7 1.7 1.7 1.7 1.8 1.8 1.8 1.8 1.7 1.8 1.8 1.8 1.8 1.7 1.7 1.7 1.6 1.7 1.7 1.7 1.6 1.7 1.7 1.7 1.7 1.7 1.7 4 SPAT02 2.3 3.2 0.0 1.7 1.7 1.7 1.7 1.7 1.8 1.8 1.8 1.8 1.7 1.8 1.8 1.8 1.8 1.7 1.7 1.7 1.6 1.7 1.7 1.7 1.6 1.7 1.7 1.7 1.7 1.7 1.7 5 CN4226 12.6 12.9 13.5 13.5 0.0 0.5 0.5 0.5 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.2 1.2 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 6 CN3853 12.6 12.9 13.5 13.5 0.0 0.5 0.5 0.5 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.2 1.2 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 7 S1755 12.3 12.6 13.5 13.5 0.9 0.9 0.0 0.0 1.4 1.4 1.3 1.3 1.3 1.3 1.4 1.3 1.3 1.3 1.3 1.3 1.2 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 8 CN677 12.3 12.6 13.5 13.5 0.9 0.9 0.0 0.0 1.4 1.4 1.3 1.3 1.3 1.3 1.4 1.3 1.3 1.3 1.3 1.3 1.2 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 9 S1165 12.3 12.6 13.5 13.5 0.9 0.9 0.0 0.0 1.4 1.4 1.3 1.3 1.3 1.3 1.4 1.3 1.3 1.3 1.3 1.3 1.2 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 10 S7643 14.0 14.0 14.3 14.3 7.0 7.0 7.3 7.3 7.3 0.0 0.9 0.8 0.8 0.9 0.9 0.9 0.9 1.1 1.1 1.1 1.1 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 11 AO27 14.0 14.0 14.3 14.3 7.0 7.0 7.3 7.3 7.3 0.0 0.9 0.8 0.8 0.9 0.9 0.9 0.9 1.1 1.1 1.1 1.1 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 12 CN686 12.6 12.6 13.5 13.5 6.4 6.4 6.7 6.7 6.7 3.2 3.2 0.6 0.5 0.6 0.6 0.7 0.7 1.2 1.2 1.2 1.1 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 13 S7161 13.2 13.2 13.5 13.5 6.4 6.4 6.7 6.7 6.7 2.9 2.9 1.2 0.3 0.4 0.4 0.5 0.4 1.2 1.2 1.1 1.1 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 14 S7150 12.9 12.9 13.2 13.2 6.1 6.1 6.4 6.4 6.4 2.9 2.9 0.9 0.3 0.3 0.4 0.5 0.4 1.2 1.2 1.1 1.1 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 15 CN3750 12.6 12.6 13.5 13.5 5.8 5.8 6.1 6.1 6.1 3.2 3.2 1.2 0.6 0.3 0.5 0.5 0.5 1.2 1.2 1.2 1.1 1.4 1.4 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 16 OM100 13.5 13.5 13.7 13.7 6.7 6.7 7.0 7.0 7.0 3.2 3.2 1.5 0.9 0.6 0.9 0.5 0.4 1.2 1.2 1.2 1.2 1.4 1.4 1.3 1.4 1.3 1.3 1.3 1.3 1.3 1.3 17 OM66 13.7 13.7 14.0 14.0 6.7 6.7 7.0 7.0 7.0 3.2 3.2 1.8 1.2 0.9 1.2 0.9 0.3 1.2 1.1 1.1 1.1 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 18 OM99 13.5 13.5 13.7 13.7 6.7 6.7 7.0 7.0 7.0 2.9 2.9 1.5 0.9 0.6 0.9 0.6 0.3 1.2 1.1 1.1 1.1 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 1.3 19 CN3989 12.6 12.6 12.6 12.6 6.7 6.7 7.0 7.0 7.0 5.3 5.3 5.8 5.6 5.6 5.8 6.1 5.6 5.6 0.6 0.6 0.6 0.8 0.8 0.7 0.8 0.7 0.7 0.7 0.7 0.7 0.7 20 CN664 13.2 13.2 12.6 12.6 6.1 6.1 6.4 6.4 6.4 5.3 5.3 5.8 5.6 5.6 5.8 6.1 5.6 5.6 1.5 0.3 0.4 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 21 CN2575 12.9 12.9 12.3 12.3 5.8 5.8 6.1 6.1 6.1 5.0 5.0 5.6 5.3 5.3 5.6 5.8 5.3 5.3 1.2 0.3 0.3 0.8 0.8 0.7 0.7 0.7 0.7 0.7 0.7 0.7 0.7 22 CN681 13.2 13.2 12.6 12.6 5.6 5.6 5.8 5.8 5.8 4.7 4.7 5.3 5.0 5.0 5.3 5.6 5.0 5.0 1.5 0.6 0.3 0.8 0.8 0.7 0.8 0.7 0.7 0.7 0.7 0.7 0.7 23 CN2641 12.6 12.6 12.6 12.6 6.4 6.4 6.7 6.7 6.7 6.7 6.7 7.0 6.7 6.7 7.0 7.3 6.7 6.7 2.3 2.3 2.0 2.3 0.0 0.4 0.4 0.3 0.3 0.3 0.3 0.3 0.3 24 CN3433 12.6 12.6 12.6 12.6 6.4 6.4 6.7 6.7 6.7 6.7 6.7 7.0 6.7 6.7 7.0 7.3 6.7 6.7 2.3 2.3 2.0 2.3 0.0 0.4 0.4 0.3 0.3 0.3 0.3 0.3 0.3 25 CN3970 12.6 12.6 12.6 12.6 6.1 6.1 6.4 6.4 6.4 6.4 6.4 6.7 6.4 6.4 6.7 7.0 6.4 6.4 2.0 2.0 1.8 2.0 0.6 0.6 0.4 0.3 0.3 0.3 0.3 0.3 0.3 26 CN8706 12.0 12.0 12.0 12.0 6.4 6.4 6.7 6.7 6.7 6.7 6.7 7.0 6.7 6.7 7.0 7.3 6.7 6.7 2.3 2.3 2.0 2.3 0.6 0.6 0.6 0.3 0.3 0.3 0.3 0.3 0.3 27 CN8281 12.3 12.3 12.3 12.3 6.1 6.1 6.4 6.4 6.4 6.4 6.4 6.7 6.4 6.4 6.7 7.0 6.4 6.4 2.0 2.0 1.8 2.0 0.3 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0 28 CN3986 12.3 12.3 12.3 12.3 6.1 6.1 6.4 6.4 6.4 6.4 6.4 6.7 6.4 6.4 6.7 7.0 6.4 6.4 2.0 2.0 1.8 2.0 0.3 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0 29 CN8355 12.3 12.3 12.3 12.3 6.1 6.1 6.4 6.4 6.4 6.4 6.4 6.7 6.4 6.4 6.7 7.0 6.4 6.4 2.0 2.0 1.8 2.0 0.3 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0 30 CN7658 12.3 12.3 12.3 12.3 6.1 6.1 6.4 6.4 6.4 6.4 6.4 6.7 6.4 6.4 6.7 7.0 6.4 6.4 2.0 2.0 1.8 2.0 0.3 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0 31 CN3412 12.3 12.3 12.3 12.3 6.1 6.1 6.4 6.4 6.4 6.4 6.4 6.7 6.4 6.4 6.7 7.0 6.4 6.4 2.0 2.0 1.8 2.0 0.3 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0 32 CN7819 12.3 12.3 12.3 12.3 6.1 6.1 6.4 6.4 6.4 6.4 6.4 6.7 6.4 6.4 6.7 7.0 6.4 6.4 2.0 2.0 1.8 2.0 0.3 0.3 0.3 0.3 0.0 0.0 0.0 0.0 0.0